Sub-dictionary based sparse representation for efficient super-resolution image reconstruction

Super-resolution image reconstruction is an important digital image processing technique, which can improve the visual effects of images or serve as a pre-processing technique. Because of its impressive reconstruction results, sparse representation based super-resolution image reconstruction has become the focus of recent research. In order to alleviate the high computational complexity of the traditional sparse representation schemes, this study presents a fast sub-dictionary-based super-resolution reconstruction method. For each small input image block, a sub-dictionary is adaptively selected and thus the high-dimensional redundant dictionary-based sparse representation vector is replaced by a low-dimensional sub-dictionary based representation vector, the computational complexity is therefore reduced Experimental results demonstrate that the proposed method can enhance the visual effects of images with a significantly low computational complexity.